The influence of the span/depth ratio when measuring the mode II fracture toughness of wood by endnotched flexure (ENF) tests was examined. Western hemlock (Tsuga heterophylla Sarg.) was used for the specimens. The ENF tests were conducted by varying the span/depth ratios; and the fracture toughness at the beginning of crack propagation GIIc was calculated by two equations that require the load-deflection compliance or Young's modulus. Additionally, the influence of the span/ depth ratio on the load-deflection compliance was analyzed by Timoshenko's bending theory in which additional deflection caused by the shearing force is taken into account. The following results were obtained: (1) When the span/depth ratio was small, the fracture toughness calculated with the data of load-deflection compliance was large. In contrast, the fracture toughness calculated with the equation containing Young's modulus tended to be constant. (2) In the small span/depth ratio range, the load-deflection compliance was estimated to be larger than that predicted by Timoshenko's bending theory. (3) To obtain the proper fracture toughness of wood with a single load-deflection relation, the span/depth ratio should be larger than that determined in several standards for the simple bending test method of wood, 12:16.